Some quick results based on GeForce 7900GTX (650/800), with stock cooler at max speed. Note that test results were obtained immediately upon applying after market thermal pastes. I will update these results after 7 days as per recommended by Arctic Silver & Zalman. Tests were conducted on an open test bench with room temperature recorded at 28ºC throughout the test. GPU was loaded using ATITool's "Show 3D View".

Stock TIM
GPU Core Idle: 45ºC
GPU Core Load: 62ºC

Arctic Silver 5
GPU Core Idle: 45ºC
GPU Core Load: 63ºC

Zalman STG1
GPU Core Idle: 45ºC
GPU Core Load: 67ºC

As you can see, the STG1 did not impress me that much... at least during the first day of testing. Right after application, the AS5 is easily better than STG1. When compared to the already hardened stock TIM, AS5 was just 1ºC behind it, whereas the STG1 was 5ºC behind.

During my testing, I found that the STG1 to be quite watery when compared to AS5, Ceramique, stock TIM or any other TIM I've tested. So, I hope to see better results once the TIM has settled in. For the record, I double checked and reapplied the STG1 on the GPU core just to make sure that the TIM spreads evenly.

Thanks to Moderno for providing the sample. Look forward for a comprehensive review on my website soon.

This post has been edited by pcmoddingmy: Dec 2 2006, 12:41 AM

As much as I would like to compare it, its sad that I don't have the Nano with me during the time of review.
Btw, next time don't advertise in your post. Move it to your signature...

Yup... because NVIDIA used ShinEtsu thermal paste I guess. Its quite thick, hard and silver based too. AS5 did quite well, even without properly curing. If any of the TIMs could do at least 1ºC better than the stock TIM, it's already pretty damn good.

This post has been edited by pcmoddingmy: Dec 2 2006, 02:17 PM

Yup... I will retest the STG1 in a the next few days.
As for AS5, I will have to reapply it and wait another 1 week... darn time consuming... LOL. Arctic Silver said it needs 200 hrs (about 8 days) to properly cure in.

Yup.. As stated on STG1's package "After application, approximately 7 days of regular computer use will give optimal performance"

Yup... provided your CPU runs at 60ºC or higher during the curing process. It will cure faster, but also risk killing your proc quicker as you need to do it about 3-4 days straight.
So better play safe and let it slowly cure in as per recommended.

Nice results there but having 8ºC drop in GPU temp by just changing thermal paste sounds ridiculous. Don't get me wrong but you could be mislead by your setup.

Firstly, I see that you are using dual 7300GT in SLI mode. So, there should be two GPU core temperatures displayed in NVIDIA Control Panel. Since you are only showing a single GPU temperature, I'm not sure if its GPU1 or GPU2.

Secondly, I see that you have a Delta fan blowing towards the GPUs. Not sure if the fan was there during the initial test w/o the STG1. If the Delta fan was there from the begining, it should be in the exact same position. If the fan moved a little, the fan could be blowing towards another direction from where it was originally seated.

Thirdly, you didn't provide any ambient/room temperatures for your setup. Room temp could have dropped after the initial testing w/o the STG1. You also didn't state the stres (100% load) or idle duration for both CPU/GPU. You have to make sure the testing period is consistent.

Lastly, I noticed that your GPU temp (with STG1) did not increase when overclocking. At 500MHz, its loaded at 49C, and maintained 49C even when overclocked to 620MHz. I suspect the GPU could have reverted back to stock clock instead of staying overclocked. This is why you should have RivaTuner's Hardware Monitoring tool running throughout the test.

I hope you don't get me wrong here but we appreciate your effort in doing these tests. Its just these few little things that you need to consider before doing such tests. On the other hand, your CPU results seems to be pretty consistent.
And btw, the original thermal paste on the 7300GT is not by NVIDIA... its applied by Galaxy. NVIDIA doesn't sell 7300GT with Zalman coolers.

Actually you no need to be pro lah.
I'm sure even normal user will feel weird when they see the GPU temp at 500MHz = 49C and overclocked to 620MHz also = 49C. Besides, having 8C drop by just changing thermal paste is a miracle, unless the original thermal paste was insufficient.

bulibulizaimon, can you please check again to see if the GPU temp is still maintained at 49C even when overclocked 620MHz? This time, have RivaTuner Hardware Monitoring tool runing okay.

I'm seeing a few users here experiencing higher temperature after applying STG1 (before cure in), which seems to be similar to my experience. For some reasons, you happen to be the only one to obtain better results after applying STG1.

Okay... just tested my 7900GTX again after about 7 days.
The original result was 67°C on load but now I'm getting about 0.5°C to 1°C drop on load.
I ripped open the cooler and found that the STG1 paste had already hardened, which is a good sign.

Since some users here suggested that I need to apply more TIM, I thought I might just give it another try. So, I reapplied some more paste (3rd application) and I tried to apply as much as I could. Hooked it up again, temps went back to 67-68°C. About 1°C higher that my first test!
So, took it out, checked the TIM and it flooded the GPU core area. I knew this would happen but it didn't help to improve the load temperature. So, applying more STG1 doesn't help to improve results.

Then, I cleaned up the GPU core and applied some generic "white" TIM on the GPU. Now I'm getting about 68°C on the white TIM, which is about the same as STG1.

Basically, I don't think the STG1 is any better than the Arctic Silver 5. I gave it 7 days to cure in but it still did not impress me. AS5 was easily 3°C better than the STG1.

AS5 is cheaper and it performs better, no doubt about it. It takes lesser time to cure in and you can expect good performance on the first day itself. Also, its easier for me just to apply on the core instead of applying on both surfaces as required by the SGT1. So indirectly, you will end up using more thermal paste with SGT1 than with AS5.

However, if you're a Zalman fan or feel like giving it a try, you should check it out. The bottle size is about the same size of an AS5 (3.5g), so don't expect it to be a biggie... Peace out~

Update: Guys, I've added a few more comments in my post. Please take note...

This post has been edited by pcmoddingmy: Dec 8 2006, 02:20 PM

Talking about cleaning heatsinks, guess what I use to clean heatsink base and CPU IHS...

I use minyak kapak..... Yes, minyak kapak (Axe Branded Medicated Oil).
Just put one or two drops onto a tissue paper and wipe it over the heatsink base or CPU IHS top. It cleans pretty well but leaves a little oil on the surface. So, grab another clean tissue and wipe away the oil. Go give it a try and restore that shine back to your heatsink/CPU!

Minyak kapak good for nose blocks, headaches, etc and now cleaning heatsinks too! No kidding~



This post has been edited by pcmoddingmy: Dec 12 2006, 02:06 AM

Because minyak kapak contains methyl salicylate, which is a sort of liquid ester. Ester is from an organic acid and an alcohol. So, this makes the minyak kapak carry certain percentage of alcohol.

Not really, coz I just use one or two drops only. IIRC, Arctic Cleaner has stronger smell right?

Well, if it cleans well... I don't see any reason why it could affect the TIM performance.

This post has been edited by pcmoddingmy: Dec 12 2006, 03:34 PM

Won't leave any noticeable residue, unless you pour the whole bottle onto the heatsink.
One drop, or even a little dip onto a tissue is more than enough. It cleans pretty well, compared to just using tissue paper to wipe the base.
Anyways, no need to worry about oil because it doesn't stick to heatsinks. Can easily be wiped away. Besides, minyak kapak is not as oily as "real oil". It has less than 30% oil on it.
You guys should really try it out. It works for me, that's what that matters. I even bought a small bottle for the purpose of cleaning heatsinks and anything that has thermal paste on it

Actually, come to think of it, the "rice grain" method only works well with exposed cores (without IHS). Check out Arctic Silver's latest instruction on applying Arctic Silver 5 on Conroe (or other similar processors):


This is known as "thin line" or "srting" method.

Might work with other type of TIM, provided its quite thick.
Trial and error is your best friend!

Yes, the "string" method applies for dual core processors such as Conroe, PressHoT, X2 and FX62.
"Rice" method is for single core and exposed cores.

Anyways, I still don't think getting a credit card to flat the TIM is a good idea. Why not let the heatsink flatout the TIM?